Question Answering on CommonsenseQA (PR-AUC)
0.595PR-AUCDissimilarity
Evaluation Results
| Method | Links | |
|---|---|---|
| DissimilarityStrategy=beam-search, Backbone=Gemma 3 4B base2025.12 | 0.595 | |
| EccentricityStrategy=beam-search, Backbone=Gemma 3 4B base2025.12 | 0.59 | |
| CocoaMSPStrategy=beam-search, Backbone=Gemma 3 4B base2025.12 | 0.586 | |
| CocoaMSPStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.12 | 0.583 | |
| CocoaPPLStrategy=beam-search, Backbone=Gemma 3 4B base2025.12 | 0.582 | |
| CocoaPPLStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.12 | 0.58 | |
| EigVecDissimilarityStrategy=beam-search, Backbone=Gemma 3 4B base2025.12 | 0.562 | |
| DissimilarityStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.12 | 0.561 | |
| EccentricityStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.12 | 0.549 | |
| EigVecDissimilarityStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.12 | 0.512 | |
| Degree MatrixStrategy=beam-search, Backbone=Gemma 3 4B base2025.12 | 0.444 | |
| Semantic EntropyStrategy=beam-search, Backbone=Gemma 3 4B base2025.12 | 0.436 | |
| Degree MatrixStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.12 | 0.409 | |
| Semantic EntropyStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.12 | 0.407 | |
| SARStrategy=beam-search, Backbone=Gemma 3 4B base2025.12 | 0.209 | |
| SARStrategy=multinomial-sampling, Backbone=Gemma 3 4B base2025.12 | 0.183 |